The aims of this proposal are to (1) improve comparative modeling so that atomic level accuracy models can be routinely generated starting from structures of homologous proteins, (2) transform the process of NMR structure determination by making possible the determination of atomic level accuracy models without side chain assignments, and (3) enable the determination of high resolution structures from 3-4.5 E resolution density maps. These methodological advances will be extended to enable structure determination of membrane proteins and large homo-oligomers based on datasets not sufficient to allow structure determination using conventional methods. PUBLIC HEALTH RELEVANCE: The proposed work could transform the process of macromolecular structure determination, making it possible to determine accurate structures rapidly with less cost and effort, and to determine atomic level structures for proteins and complexes for which this is not currently possible. Such structures could provide insight into fundamental biological processes and the basis for disease causing mutations, and allow the structure based design of new therapeutics.